Ayasdi

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search
Symphony AyasdiAI [1]
Private
IndustryEnterprise software
Founded2008
HeadquartersRedwood City, California,
Key people
Simon Moss
(CEO)
Peter Downs
(CFO)
Richard Stocks
(CPO)
Stephen Moss
(Head of Customer Success)
ProductsAML, Clinical Variation Management, Program Performance Intelligence, Denials Management
ServicesFinancial Services FinTech
Number of employees
100+ (2020)
ParentSymphony AI
WebsiteAyasdi

Symphony AyasdiAI is a machine intelligence software company that offers a software platform and applications to organizations looking to analyze and build predictive models using big data or highly dimensional data sets. Organizations and governments have deployed Ayasdi's software across a variety of use cases including the development of clinical pathways for hospitals,[2] anti-money laundering, fraud detection, trading strategies, customer segmentation, oil and gas well development, drug development, disease research, information security, anomaly detection, and national security applications.[3][4]


Ayasdi focuses on unsupervised machine learning at scale.[5] In effect, the Ayasdi system consumes the target data set, runs many different unsupervised and supervised machine learning algorithms on the data, automatically finds and ranks best fits, and then applies topological data analysis to find similar groups within the resultant data. It presents the end analysis in the form of a network similarity map, which is useful for an analyst to use to further explore the groupings and correlations that the system has uncovered. This reduces the risk of bias since the system surfaces "what the data says" in an unbiased fashion, rather than relying on analysts or data scientists manually running algorithms in support of pre-existing hypotheses.[6] Ayasdi then generates mathematical models which are deployed in predictive and operational systems and applications.


Organizations using Ayasdi have found Ayasdi's automated, platform-based approach to machine intelligence to be two to five orders of magnitude more efficient than existing approaches to big data analytics, as measured in the amount of time and expense required to complete analysis and build models using large and complex data sets. One widely reported example at a top five global systemically important bank was that to build models required for the annual Comprehensive Capital Analysis and Review (CCAR) process took 1,800 person-months with traditional manual big data analytics and machine learning tools, but took 6 person-months with Ayasdi. A project at a second global systemically important bank showed Ayasdi reducing the time to build risk models from 3,000 person-hours to 10 minutes.[7]

History and funding[edit]

Ayasdi was founded in 2008 by Gunnar Carlsson, Gurjeet Singh, and Harlan Sexton after 12 years of research and development at Stanford University.[3][4] While at Stanford, the founders received $1.25 million in DARPA and IARPA grants for "high-risk, high-payoff research".[3] In 2012 Ayasdi landed a Series A round of funding led by Floodgate Capital and Khosla Ventures for $10.25 million.[8] On July 16, 2013, Ayasdi closed $30.6 million in Series B funding from Institutional Venture Partners, GE Ventures, and Citi Ventures.[9] On March 25, 2015, Ayasdi announced a new $55 million round of Series C funding, led by Kleiner Perkins Caufield & Byers, and joined by four current investors, Institutional Venture Partners, Khosla Ventures, Floodgate Capital, Citi Ventures, and two new investors, Centerview Capital Technology and Draper Nexus.[10]


On May 15th, 2019 Ayasdi was acquired by SymphonyAI (SAI) group and renamed as Symphony AyasdiAI. SAI was founded by Dr. Romesh Wadhwani and is a $1 billion investment platform focused on building the next generation of artificial intelligence and machine learning companies in Healthcare, Financial Services, Retail, Industrial, Media and Enterprise markets.[11]

Product[edit]

Ayasdi is a machine intelligence platform. It includes dozens of statistical and both supervised and unsupervised machine learning algorithms and can be extended to include whatever algorithms are required for a particular class of analysis. The platform is extensively automated and is in production at scale at many global 100 companies and at governments in the world. It features Topological Data Analysis as a unifying analytical framework, which automatically calculates groupings and similarity across large and highly dimensional data sets, generating network maps that greatly assist analysts in understanding how data clusters and which variables are relevant. When compared with manual approaches to statistical analysis and machine learning, results with Ayasdi will typically be achieved much faster to achieve and more accurate due to the automation and scalability built into the platform. The Ayasdi platform also develops mathematical models, including predictive models, based on the results of the analysis. This allows Ayasdi to deployed as an operational system, or as a part of operational systems, and not just for analysis.[12]


In September 2019, Ayasdi announced the launch of Ayasdi AML, an anti-money laundering solution to combat fraudulent money laundering schemes.[13] Ayasdi AML works with existing data to provide end-to-end intelligence spanning transaction management systems (TMS), event dispositioning, and case management systems. Ayasdi AML uses a new approach (intelligent segmentation scheme) that is used within the TMS to reduce false alerts and are human-explainable to facilitate rapid regulatory approval. Additionally, Ayasdi AML surfaces hidden relationships that exist and identifies meaningful ones; monitors TMS-generated events and intelligently dispositions them; and accelerates investigations by providing a 360⁰ view of the customer, behavioral transitions, and more. Ayasdi AML applies groundbreaking machine learning to existing transactional data, generating new insights and accurate alerts. Symphony AyasdiAI’s powerful topological data analysis (TDA) creates intelligent segments of customers based on patterns of behavior. This results in fewer false positives, enabling you to proactively detect, investigate, and report suspicious activities. [14]


Liquidity Optimization – Predict an optimal liquidity level adhering to operational and regularity requirements using Ayasdi Liquidity Optimization. Automatically predict optimal cash balances for each client. [15]


KYC – Perform detailed dynamic analysis of your customers to determine behavioral changes since the last review or onboard. Unsupervised machine learning produces a detailed analysis to create new trigger events for review. [16]

Applications[edit]

Clinical Variation Management from AyasdiAI is a cornerstone application for a broader clinical transformation that will touch every part of the delivery chain – from how hospitals are organized, the conversations they have with their doctors, the confidence with which they take on risk-sharing arrangements with payers, and most importantly the care they deliver to their patients. The key lies in finding the patterns and relationships that matter in complex, multi-structured clinical, and billing data. [17] Flager hospital was able to discover a care path for pneumonia using AyasdiAI CVM which offered superior care: • Their length of stay was two days shorter • Their cost per episode was more than $1.3K less expensive • Their readmissions rates were nearly 7X lower [18]

AyasdiAI’s Clinical Variation Management application: ● Discovers what’s going on in the hospital ● Surfaces the best care practices ● Builds new care paths to be added to coordination systems ● Continuously improves the care paths [19]


AyasdiAI Fraud Waste and Abuse App rapidly correlates and analyzes thousands of attributes simultaneously and groups data points that are similar to reveal patterns and outliers through visual networks. The software automatically lists the statistically significant features that characterize these patterns and outliers. These features can then be used to develop more effective, localized fraud detection rules and models. AyasdiAI’s machine intelligence software helps payers and/or providers achieve the following: ● Identify new patterns of aberrant behavior ● Validate and improve existing detection models ● Prioritize fraud leads for SIU teams ● Improve detection in the pre-payment cycle [20]


AyasdiAI’s Program Performance Intelligence application looks at each component in the program review process, quantitative, qualitative, and exogenous, compares that with data on the progression of other programs, and makes an unbiased assessment of the future of the program. To achieve this, Ayasdi has created a Topological Data Analysis-powered application to evaluate the available data, surfacing patterns, and relationships that indicate program performance changes. Using these relationships, the application predicts the future health of the program and details exactly what is driving those findings in a simple to understand terms. Furthermore, the application learns constantly and takes advantage of each new data point to refine its thinking. [21]

Users and industries[edit]

Ayasdi customers include many large enterprises, hospitals, medical research institutions, and governments across industries including financial services, healthcare, manufacturing, security, life sciences, and the public sector.[22][23]

References[edit]

  1. ^ https://www.ayasdi.com/
  2. ^ "Intermountain to deploy clinical variation management software from Ayasdi". Health Care IT News. March 2, 2016. Retrieved March 2, 2016.
  3. ^ a b c "Ayasdi: A Big Data Start-Up With a Long History". The New York Times. January 16, 2013. Retrieved March 5, 2013.
  4. ^ a b "A cure for cancer? This 'big data' startup says it can deliver". Venturebeat. January 16, 2013. Retrieved March 5, 2013.
  5. ^ "Knowing What's Possible a Big Obstacle for Big Data". Datanami. February 1, 2016. Retrieved February 1, 2016.
  6. ^ "How a 'Nuisance Variable' Turned Into Potential Lifesaver". Datanami. January 4, 2016. Retrieved January 4, 2016.
  7. ^ https://www.ayasdi.com/blog/bigdata/yesterday-ccar-less-stressful-citigroup/
  8. ^ "Venture capital deals". CNNMoney. January 16, 2013. Archived from the original on 2013-03-12. Retrieved March 5, 2013.
  9. ^ "News and Events - Ayasdi". Ayasdi.com. Retrieved July 3, 2017.
  10. ^ "News and Events - Ayasdi". Ayasdi.com. Retrieved July 3, 2017.
  11. ^ https://www.ayasdi.com/company/news-and-events/press/ayasdi-joins-symphonyai-portfolio/
  12. ^ Erin Bury (January 16, 2013). "BetaKit » Ayasdi Comes Out of Stealth With $10.25M to Answer Unknown Data Questions". Betakit. Archived from the original on 2013-03-08. Retrieved July 3, 2017.
  13. ^ "Symphony AyasdiAI Launches Next-Generation AI Solution for Anti-Money Laundering". www.businesswire.com. 2019-09-24. Retrieved 2020-03-05.
  14. ^ https://s3.amazonaws.com/cdn.ayasdi.com/wp-content/uploads/2019/09/16124946/Ayasdi-AML-Data-Sheet.pdf
  15. ^ https://s3.amazonaws.com/cdn.ayasdi.com/wp-content/uploads/2018/12/23132421/Ayasdi-Financial-Services-Use-Cases.pdf
  16. ^ https://s3.amazonaws.com/cdn.ayasdi.com/wp-content/uploads/2018/12/23132421/Ayasdi-Financial-Services-Use-Cases.pdf
  17. ^ https://s3.amazonaws.com/cdn.ayasdi.com/wp-content/uploads/2018/01/07233139/CS-Flagler-07.24.18.pdf
  18. ^ https://www.hfma.org/topics/financial-sustainability/article/using-artificial-intelligence-enabled-flagler-hospital-reduce-clinical-variation.html
  19. ^ https://s3.amazonaws.com/cdn.ayasdi.com/wp-content/uploads/2018/12/21221934/WP-Ayasdi-Healthcare-IntelligentApps-08.22.2018-V2.pdf
  20. ^ https://s3.amazonaws.com/cdn.ayasdi.com/wp-content/uploads/2018/12/21221934/WP-Ayasdi-Healthcare-IntelligentApps-08.22.2018-V2.pdf
  21. ^ https://s3.amazonaws.com/cdn.ayasdi.com/wp-content/uploads/2018/06/07233655/WP-Ayasdi-Program-Performance-Intelligence.pdf
  22. ^ "DARPA-Backed Ayasdi Launches With $10M From Khosla, Floodgate To Uncover The Hidden Value In Big Data". Techcrunch. January 16, 2013. Retrieved March 5, 2013.
  23. ^ "Extracting insights from the shape of complex data using topology". Nature. September 13, 2012. Retrieved April 1, 2013.

External links[edit]